Screening Newborns for Congenital Cytomegalovirus Infection

Abstract

IN THIS ISSUE OF JAMA, BOPPANA AND COLLEAGUES 1 DEscribe their attempts to validate universal screening methods for congenital cytomegalovirus (CMV) infection, the most common congenital viral infection in the United States and many other regions, by using polymerase chain reaction (PCR) analysis of newborn dried blood spots. Congenital CMV infection remains an important public health problem not only because it has accounted for as much or more disability over the past 50 years than was associated with congenital rubella syndrome, but also because CMV infection, whether symptomatic or silent at birth, represents the most common nongenetic cause of permanent hearing loss among children in the United States. Congenital CMV infection causes a substantial proportion of sensorineural hearing loss even in regions with a low prevalence. Culture surveillance of nearly 4000 newborns in Utah revealed a congenital infection rate of 0.29%, corresponding to approximately 3 infected infants for every 1000 newborns or 165 infected newborns annually statewide, given the state’s current birth rate. Because as many as 10% of these infants will have CMV disease, a condition associated with a 30% to 50% prevalence of sensorineural hearing loss, and the remainder will have silent CMV infection, a condition conservatively associated with a 6% to 8% rate of sensorineural hearing loss, CMV accounts for hearing loss in 15 to 20 children each year in Utah, or approximately 20% of the 90 to 120 Utah children identified with permanent hearing loss annually (J. C. Carey, University of Utah School of Medicine, written communication, March 2010). Worldwide, thousands of infants have CMV-induced hearing loss annually. Universal newborn screening for many disorders has immense public health benefits. The history of phenylketonuria (PKU) illustrates the enormous benefit of such strategies. Universal screening of newborn hearing has led to improvement in language outcomes, with greater improvements when deaf infants are identified earlier and interventions are begun sooner. Thus, it seems logical to infer that if CMV-infected infants can be identified in the neonatal period, hearing could be tracked closely, interventions could begin early, and the hearing outcomes of congenital CMV infection might be improved. However, 2 challenges confront universal screening for congenital CMV infection. First, universal screening is only justified when an effective intervention exists that prevents adverse health outcomes, such as averting intellectual disability when infants who lack phenylalanine hydroxylase initiate a diet low in phenylalanine early in life. At present, no such solution exists for CMV. Although neonatal ganciclovir therapy improves hearing outcomes in some infants, the benefits are not universal and the risk/benefit ratio of ganciclovir therapy in infants with silent congenital CMV infection has not been sufficiently well-established to recommend such therapy for all CMV-infected infants. Nonetheless, knowing which infants are infected with CMV has benefit through early identification of infants at risk of CMVinduced hearing loss and may have additional benefits in the future when new and more effective antiviral strategies become available. The second challenge, the one addressed by Boppana et al, is the availability of a highly sensitive and specific screening method. Neonatal screening of hearing alone is insufficient to detect all CMV-infected infants because most infants with congenital CMV infection have normal hearing in the neonatal period. Fowler et al demonstrated that nearly 20% of the children with hearing loss due to silent congenital CMV infection have delayed onset of sensorineural hearing loss. In addition, most infants with congenital hearing loss do not have CMV infection. In 1996, when Barbi et al demonstrated that newborn dried blood spots and PCR assays could be used to detect CMV DNA in the blood of congenitally infected infants, there was optimism that the opportunity for universal screening for CMV had arrived. However, the study by Barbi et al was not population-based and did not have a sufficient sample size to establish convincingly the sensitivity and specificity of mass screening of newborn dried blood spots. In contrast, the large, prospective study by Boppana et al shows that even though real-time PCR has excellent specificity (approaching 100%), the sensitivity of the PCR